Systematic Review of Herbs and Dietary Supplements for Glycemic Control in Diabetes

Reviews/Commentaries/Position Statements

Systematic Review of Herbs and Dietary
Supplements for Glycemic Control in
Diabetes

GLORIA Y. YEH, MD, MPH1,2 TED J. KAPTCHUK, OMD1 (CAM) among the general public (2,3),
DAVID M. EISENBERG, MD1 RUSSELL S. PHILLIPS, MD1,2 the American Diabetes Association issued
a Position Statement in 2001 on “Unprov-
OBJECTIVE — To conduct a systematic review of the published literature on the efficacy and en Therapies” that encouraged health care
safety of herbal therapies and vitamin/mineral supplements for glucose control in patients with providers to ask their patients about alter-
diabetes. native therapies and practices, evaluate
each therapy’s effectiveness, be cognizant
RESEARCH DESIGN AND METHODS — We conducted an electronic literature search of any potential harm to patients, and ac-
of MEDLINE, OLDMEDLINE, Cochrane Library Database, and HealthSTAR, from database knowledge circumstances in which new
inception to May 2002, in addition to performing hand searches and consulting with experts in and innovative diagnostic or therapeutic
the field. Available clinical studies published in the English language that used human partici- measures might be provided to patients
pants and examined glycemic control were included. Data were extracted in a standardized (4).
manner, and two independent investigators assessed methodological quality of randomized
controlled trials using the Jadad scale. Recently, two national surveys have
examined CAM use among those with di-
RESULTS — A total of 108 trials examining 36 herbs (single or in combination) and 9 vitamin/ abetes. One study, using 1996 Medical
mineral supplements, involving 4,565 patients with diabetes or impaired glucose tolerance, met Expenditure Panel Survey data, reported
the inclusion criteria and were analyzed. There were 58 controlled clinical trials involving individuals that ϳ8% of respondents with diabetes
with diabetes or impaired glucose tolerance (42 randomized and 16 nonrandomized trials). Most saw a CAM professional for care (5). A
studies involved patients with type 2 diabetes. Heterogeneity and the small number of studies per nationally representative survey conducted
supplement precluded formal meta-analyses. Of these 58 trials, the direction of the evidence for in 1997–1998 reported that about one-
improved glucose control was positive in 76% (44 of 58). Very few adverse effects were reported. third of respondents with diabetes use
CAM to treat their condition (6). In other
CONCLUSIONS — There is still insufficient evidence to draw definitive conclusions about surveys of specific diabetic populations,
the efficacy of individual herbs and supplements for diabetes; however, they appear to be 39% of Navajo, two-thirds of Vietnamese,
generally safe. The available data suggest that several supplements may warrant further study. and 49% of a largely Hispanic population
The best evidence for efficacy from adequately designed randomized controlled trials (RCTs) is in South Texas used CAM (7–9).
available for Coccinia indica and American ginseng. Chromium has been the most widely studied
supplement. Other supplements with positive preliminary results include Gymnema sylvestre, In general, the scientific literature on
Aloe vera, vanadium, Momordica charantia, and nopal. the efficacy of CAM in the treatment of
diabetes is relatively sparse and heteroge-
Diabetes Care 26:1277–1294, 2003 neous. Studies have examined mind-body
techniques, biofeedback, relaxation, qigong
D iabetes is a predominant public lar disease. With increasing rates of child- (10 –17), massage therapy, yoga, alterna-
health concern, affecting ϳ16 mil- hood and adult obesity, diabetes is likely tive dietary/lifestyle modifications (18),
lion persons in the U.S. The disease to become even more prevalent over the aromatherapy, acupuncture, and other
causes substantial morbidity, mortality, coming decade (1). systems of healing such as traditional
and long-term complications and remains Chinese medicine (TCM) (19 –30).
an important risk factor for cardiovascu- In response to the increasing use of
complementary and alternative medicine Most of the literature, however, has
focused on herbs or other dietary supple-
●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ments. This finding parallels results from
prevalence surveys that report herbal
From the 1Division for Research and Education in Complementary and Integrative Medical Therapies, remedies or other dietary supplements
Harvard Medical School, Boston, Massachusetts; and the 2Division of General Medicine and Primary Care, taken by mouth to be consistently among
Department of Medicine, Beth Israel Deaconess Medical Center, Boston, Massachusetts. the top CAM therapies used, regardless of
the sample surveyed (5,6,8,9,31).
Address correspondence and reprint requests to Gloria Y. Yeh, MD, Harvard Osher Institute, 401 Park Dr.,
Ste. 22A, Boston, MA 02215. E-mail: [email protected]. Plant derivatives with purported hy-
poglycemic properties have been used in
Received for publication 30 October 2002 and accepted in revised form 13 January 2003. folk medicine and traditional healing sys-
Additional information for this article can be found in an online appendix at http://care. tems around the world (e.g., Native
diabetesjournals.org. American Indian, Jewish [32], Chinese
Abbreviations: CAM, complementary and alternative medicine; GTT, glucose tolerance test; RCT, ran- [20], East Indian, Mexican). Many mod-
domized controlled trial; TCM, Traditional Chinese medicine. ern pharmaceuticals used in conventional
A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion
factors for many substances.

DIABETES CARE, VOLUME 26, NUMBER 4, APRIL 2003 1277

1278 Table 1—Controlled clinical trials of single herbs for glycemic control* Review of herbs/vitamins in diabetes

Herb/Supplement Reference Design Sample Intervention Control Outcomes Evidence Jadad Adverse Effects/
Allium sativum Sitprija S et al Double-blind; 2 33 Type 2; diet Placebo No change in FBG, Direction 2 Events
Garlic; 700 mg/d Placebo juice
(Garlic) (1987) parallel groups alone (preparation PPG; insulin – N/A No side effects; no
Aloe vera Bunyapraphatsara unspecified); for 4 wks Placebo juice ϩ effect on liver
Non-randomized; 76 Type 2; Decrease FBG function
Aloe vera N et al (1996) Single-blind; 2 uncontrolled Aloe vera Linn. 80% Distilled water ϩ
parallel groups on OHA juice; 1 tbsp BID Decrease FBG No effects on liver/
Artocarpus Yongchaiyudha S (prepared by Faculty Distilled water ϩ kidney function
heterophyllus et al (1996) Non-randomized; 40 Type 2; of Pharmacy, Mahidol Decrease PPG ϩ
Single-blind; 2 newly University, Thailand); Placebo herb tea N/A 1/40 ketosis (group
Asteracanthus Fernando MR et parallel groups diagnosed for 42 d (sape, Imperata Decrease PPG Ϫ not reported)
longifolia al (1991)† brasiliensis) ϩϩ
Non-randomized; 10 Type 2; no Aloe vera Linn. 80% No change in FBG, ϩϩ N/A Not reported
Bauhinia forficata Fernando MR et Open-label; diabetes juice; 1 tbsp BID Placebo tablet HgbA1C, insulin ϩ N/A Not reported
al (1991)† Crossover; medication (prepared by Faculty
DIABETES CARE, VOLUME 26, NUMBER 4, APRIL 2003 Coccinia indica Short-term of Pharmacy, Mahidol No treatment; OHA Decrease FBG, PPG ϩ 3 No side effects; no
Russo EMK et al metabolic trial 10 Type 2; no University, Thailand); effect on liver/
Coccinia indica (1990) diabetes for 42 d Bitter commercial tea Decrease FBG, PPG kidney function
Non-randomized; medication blend (similar to OHA)
Ficus carica (Fig leaf) Azad Khan AK et Open-label; Artocarpus heterophyllus; 4 No side effects; no
al (1979) Crossover; 16 Type 2; diet 200mg fresh leaves Placebo tablet Decrease PPG, effect on liver/
Ginseng Short-term and/or OHA boiled decoction; insulin kidney function
(Unspecified) Kamble SM et al metabolic trial single experimental requirement; no
(1996) 32 Type 2; dose prior to GTT change in FPG, N/A Not reported
Double-blind; uncontrolled or C peptide,
Serraclara A et al Crossover untreated Asteracanthus longifolia; HgbA1C 2 No side effects
(1998) 100mg fresh leaves
Double-blind; 2 70 Type 2; other boiled decoction; Decrease FBG, 3 No side effects
Sotaniemi EA et parallel groups medications single experimental HgbA1C
al (1995) unclear dose prior to GTT (200mg); no
Non-randomized; change in BG,
Open-label; 3 10 Type 1; diet Bauhinia forficata tea; insulin, C-
parallel groups and insulin 3g/d (1g individual tea peptide during
bags from dried GTT
Open-label; 36 Type 2; newly leaves); for 8 wks
Crossover diagnosed; diet
alone Coccinia indica leaf; 1800
Double-blind; 3 mg/d (freeze-dried
parallel groups powder from fresh
leaves in tablets); for 6
wks

Coccinia indica; 6g/d
(dried pellets from
fresh leaves); for 12
wks

Fig leaf tea; 13g/d leaf
decoction; for 4 wks

Ginseng; 100mg/d vs.
200mg/d (tablet
preparation Dansk
Droge, Copenhagen);
for 8 wks

DIABETES CARE, VOLUME 26, NUMBER 4, APRIL 2003 Ginseng (American) Vuksan V et al Single-blind; 10 Type 2; diet Ground root of American Identical placebo Decrease PPG (all ϩ 3 No side effects
Panax (2000) Multiple and/or OHA Ginseng; 3g vs 6g vs 9g capsule containing doses), no ϩ
quinquefolius crossover; capsules (Chai-Na-Ta corn flour difference ϩϩ 2 Mild insomnia (1/9)
Short-term 9 Type 2; well- Corp, British between doses or ϩϩ
Ginseng (American) metabolic trial controlled on Columbia); single Identical placebo administration ϩϩ 2 No effect on liver/
Panax diet and/or experimental dose at capsule containing times kidney function
quinquefolius Vuksan V et al Single-blind; OHA varying times prior to corn flour ϩ
(2000) Multiple GTT Decrease PPG ϩ N/A Not reported
Ginseng (American) crossover; 24 Type 2; diet Placebo capsule (given at 0, Ϫ
Panax Vuksan V et al Short-term and/or OHA Ground root of American Ϫ40min) ϩϩ N/A No side effects
quinquefolius (2001) metabolic trial Ginseng; 3g capsule No GS4 treatment
47 Type 2; all on (Chai-Na-Ta Corp, Decrease HgbA1C, N/A Not reported
Gymnema sylvestre Double-blind; OHA British Columbia); No GS4 treatment FBG; no change
Crossover single experimental insulin N/A No hyper-sensitivity
Gymnema sylvestre 64 Type 1; all on dose at varying times Distilled water reactions
Baskaran K et al Non-randomized; insulin prior to GTT Decrease FBG,
Momordica charantia (1990) Open-label; 2 Placebo injection HgbA1C, 3 No side effects; no
parallel groups 18 Type 2; newly American Ginseng (unspecified) glycosylated effect on liver/
Momordica charantia, diagnosed extract; 3g/d plasma protein, kidney function
V-insulin Shanmugasundaram Non-randomized; (standardized extract, Placebo herb tea conventional
9 DM (?6 Type 1, Chai-Na-Ta Corp, (sape, Imperata medication, urine 2 No side effects
Myrcia uniflora ERB et al Open-label; 2 3 Type 2); all British Columbia); for brasiliensis) glucose; increase
on insulin/OHA 8 wks insulin
Ocimum sanctum (1990) parallel groups stopped during Fresh spinach leaf
(Holy basil) study Gymnema sylvestre powder Decrease FBG,
Welhinda J et al Non-randomized; extract, GS4; 400 mg/d HgbA1C, Yeh and Associates
(1986) Open-label; 18 Type 2; on diet capsule; for 18–20 mos glycosylated
Crossover; and/or OHA plasma protein,
Baldwa VS et al Short-term Gymnema sylvestre insulin
(1977) metabolic trial 40 Type 2; on diet extract, GS4; 400 mg/d requirement,
and/or OHA capsule; for 2–30 mos urine glucose;
Russo EMK et al Non-randomized; increase C-
(1990) Blinding Momordica charantia peptide
unclear; 2 juice; homemade
parallel preparation (dose Decrease PPG
groups; Short- unspecified); single
term metabolic experimental dose Decrease FBG
trial prior to GTT
No change in FBG,
Double-blind; Momordica charantia HgbA1C;
Crossover vegetable insulin decrease insulin
(purified protein
1279 Agrawal P et al Single-blind; extract); single severity Decrease FBG, PPG,
(1996) Crossover dependent urine glucose
experimental dose
(subcutaneous)

Myrcia uniflora tea; 3g/d
(1g individual tea bags
from dried leaves); for
8 wks

Ocimum album fresh leaf;
2.5g powder; for 4 wks

Review of herbs/vitamins in diabetesOpuntia streptacantha Frati AC et alOpen-label;14 Type 2; dietGrilled nopal stems;400ml H2ODecrease glucose,ϩϩ1 Not reported
Crossover; and/or OHA 500g; single insulin ϩϩ N/A Not reported
1280(Nopal) (1990) Short-term (diet alone dur- experimental dose Water; broiled
metabolic trial ing study) zucchini squash Decrease FBG,
Opuntia streptacantha Frati-Munari AC Fresh nopal stems, insulin
Non-randomized; 32 Type 2; OHA broiled; 500g crude
(Nopal) et al (1988) Open-label; weight; single
Crossover; stopped during experimental dose
Short-term
metabolic trial study

Silymarin (Milk Velussi M et al Open-label; 2 60 Type 2 with Silymarin; 600mg/d No treatment Decrease FBG, ϩϩ 2 No side effects
Thistle) (1997) parallel groups cirrhosis; diet (“Legalon” formulation, mean BG, urine
and insulin IBI Lorenzini, Milan); glucose,
for 12 mos HgbA1C, fasting
insulin, insulin
requirement, C
peptide

Trigonella foenum Sharma RD et al Blinding unclear; 15 Type 2; diet Defatted fenugreek seed No treatment Decrease FBG, PPG, ϩϩ 1 Not reported
(Fenugreek) (1990) Crossover and OHA (dose powder; 100g/day in No treatment postprandial ϩϩ 1 Not reported
decreased 20% unleavened bread; for No treatment insulin, urine 1 Not reported
Trigonella foenum Sharma RD et al Blinding unclear; during study) 10 d No treatment glucose N/A No side effects
(Fenugreek) (1990) Crossover
5 Type 2; diet and Defatted fenugreek seed Decrease FBG, PPG,
Trigonella foenum Sharma RD et al Blinding unclear; OHA (dose powder; 100g/day in urine glucose,
(Fenugreek) (1990) Crossover decreased 20% unleavened bread; for insulin
during study) 20d
Trigonella foenum Madar Z et al Non-randomized; Decrease FBG, PPG, ϩϩ
(Fenugreek) (1988) Open-label; 10 Type 1; diet Defatted debitterised urine glucose; no
Crossover; and insulin fenugreek seed change body
Short-term (dose decreased powder; 100g/d in weight, insulin
metabolic trial during study) unleavened bread; for
10d Decrease PPG; no ϩ
21 Type 2
Fenugreek seed powder;
15g in water; single change in insulin
experimental dose with
meal tolerance test

*All trials are randomized unless otherwise specified in the “Design” column. Ϫ, no outcome measures positive; ϩ, at least one outcome measure positive; ϩϩ, Ͼ50% of outcome measures positive. FBG, fasting
blood glucose; PPG, postprandial glucose; OHA, oral hypoglycemic agent. †Fernando MR, Wickramasinghe N, Thabrew MI, Ariyananda PL, Karunanayake EH: Effect of Artocarpus heterophyllus and
Asteracantha longifolia on glucose tolerance in normal human subjects and in maturity-onset diabetic patients. J Ethnopharmacol 31:277–277, 1991

medicine today also have natural plant or-
igins. Among them, metformin was de-
rived from the flowering plant, Galega
officinalis (Goat’s Rue or French Lilac),
which was a common traditional remedy
for diabetes (33,34). Similarly, the use of
vitamin and mineral supplements for pri-
mary or secondary disease prevention is
of increasing interest (35).

However, there is relatively little known
regarding efficacy and safety of herb, vita-
min, or other dietary supplements for di-
abetes. Two prior reviews by Ernst et al.
(36,37) examined plants with hypoglyce-
mic activity in humans, including 22 clin-
ical trials (5 randomized controlled trials
[RCTs]). One recent systematic review on
Ayurvedic interventions was published
under the auspices of the Agency for
Healthcare Research and Quality (AHRQ)
(38). Additionally, there have been sev-
eral qualitative reviews reporting on se-
lected supplements used in diabetes
(33,35,39 – 45). To our knowledge, there
have been no comprehensive systematic
reviews incorporating vitamin/mineral
supplements, in addition to herbal prod-
ucts, for glucose control among patients
with diabetes.

Our objective was to review and sum-
marize the literature on herbal remedies
and dietary supplements for use in diabe-
tes, to propose guidelines that may aid
practitioners in advising their patients,
and to provide recommendations for fu-
ture research.

RESEARCH DESIGN AND
METHODS
We searched MEDLINE, OLDMEDLINE,
CAM-PubMed, HealthSTAR, and the Co-
chrane Library Database from 1960 to
March 2002 using the MeSH terms CAM,
alternative therapies, hypoglycemic plants,
and individual herb and supplement
names from popular sources, each
crossed with the term diabetes mellitus. In
addition, we contacted experts in the field
to identify studies, and we also hand-
searched references of key articles. We
did not include supplements made from
animal components. Fish oil supplemen-
tation, for example, has been examined in
prior meta-analyses (46,47). We also did
not include soluble fiber supplements,
which overlap considerably with dietary
fiber treatment and already play a role in
conventional diabetes nutrition advice
(48 –51).

We limited studies to those published

DIABETES CARE, VOLUME 26, NUMBER 4, APRIL 2003

DIABETES CARE, VOLUME 26, NUMBER 4, APRIL 2003 Table 2—Controlled clinical trials of combination herbs for glycemic control*

Herb/Supplement Reference Design Sample Intervention Control Outcomes Evidence
Direction Jadad Adverse Effects/Events

Traditional Chinese Vray M et al Double-blind; 4 216 Type 2; on Traditional Chinese herbs; 21 Placebo TCM Decrease FBG, decrease ϩ 3 Diarrhea (1), dry mouth
Medicine (TCM) (1995) parallel groups diet alone capsules/d (each containing capsule; PPG, with synergistic (1) TCM treatment;
herb combina- 150mg Coptis Chinensis, Placebo OHA effect of both Ϫ vertigo (1),
tion Double-blind; 12 Type 2; on diet 30mg Astragalus tablet treatments; no ϩ hypoglycemia (9)
Crossover and/or OHA membranaceus, 120mg change in insulin or OHA treatment;
Xiaoke (TCM) Hale PJ et al Lonicera Japonica) ϩ/Ϫ Oral “Ordinary tea” HgbA1C hypoglycemia (10)
(1989) Non-randomized; 148 Type 2 hypoglycemic (glibenclamide OHAϩTCM
Open-label; 2 7.5mg/d); for 3 mos OHA (glyburide) No change in HgbA1C,
SPDPA (TCM) Xiong M et al parallel groups FBG, PPG, or insulin 3 No side effects
formula (1995) Xiaoke tea; uncharacterized
herb preparation in 2.72g Decrease FBG, but no N/A No side effects
teabag (Home and Sutton, difference from
London), 4 infusions/d; for 4 control
wks

“Semen Persical Decoction for
Purgation with Addition”
(Rhubarb, Semen Persical,
Ramulus Cinnamomum,
Radix Glycyrrhizae, Radix
Scrophularie, Radix
Rehmanniae, Radix
Ophiopogonis, Radix
Astragalus); for 2 mos

Native American Ryan EA et al Single-blind; 2 40 Type 2; on Herbal tea; Unspecified amount, Placebo decoction No change in PPG, Ϫ 2 Minor gastrointestinal
herb combina- (2000) parallel groups diet, OHA, and/ Populus tremuloides with Chinese fructosamine, discomfort (1)
tion or insulin (trembling aspen) and green tea, HgbA1C
Heracleum lanatum (cow mint, fennel
parsnip) decoction; 250mL/ seed
d; for 10 d

Tibetan Medicine Namdul T et al Open-label; 2 200 Type 2; newly Tibetan medicine herbs; No herb Decrease FPG, PPG, ϩϩ 2 Not reported
herb combina- (2001) parallel groups diagnosed or individualized powder/pill treatment and GHb; no change
tion untreated; on combination (at least 2 of in weight
diet alone 4:Kyura-6, Aru-18, Yungwa-
4, Sugmel-19); for 6 mos

*All trials are randomized unless otherwise specified in the “Design” column. Ϫ, no outcome measures positive; ϩ, at least one outcome measure positive; ϩϩ, Ͼ50% of outcome measures positive. FBG, fasting
blood glucose; PPG, postprandial glucose; OHA, oral hypoglycemic agent.

Yeh and Associates

in the English language and restricted our
search to herbs and supplements for gly-
cemic control and symptoms of hypergly-
cemia. We excluded trials that primarily
examined diabetic complications such as
neuropathy, nephropathy, or retinopa-
thy. We included studies in subjects with
impaired glucose tolerance or those spe-
cifically at risk for diabetes (e.g., older,
sedentary, obese individuals with a family
history of diabetes). As supporting evi-
dence, we also examined studies of glyce-
mic control in healthy volunteers. To
assess quality of RCTs, we employed the
Jadad scale, a previously validated instru-
ment that assesses trials based on appro-
priate randomization, blinding, and
description of study withdrawals or drop-
outs (52,53). Quality of evidence for spe-
cific herbs or supplements was assessed
using the U.S. Preventive Services Task
Force criteria (54) (online appendix A;
http://care.diabetesjournals.org) and the
American Diabetes Association evidence
grading system for clinical practice rec-
ommendations (55) (online appendix B).
Clinical guidelines were based on the cri-
teria developed by Weiger et al. (56) (on-
line appendix C). These criteria place
individual CAM therapies along a contin-
uum that encompasses “recommend”
(high-quality evidence supports both ef-
ficacy and safety), “accept/consider rec-
ommending” (evidence supports both
efficacy and safety), “accept” (evidence re-
garding efficacy is inconclusive but sup-
ports safety), and “discourage” (evidence
indicates either inefficacy or serious risk).

Data synthesis
A total of 108 clinical studies were found
examining 25 single herbs, 11 combina-
tion herb formulas, and 9 vitamin/
mineral supplements as potential therapy
for diabetes. Of these, 58 were controlled
clinical trials in patients with diabetes or
impaired glucose tolerance (42 random-
ized, 16 nonrandomized). Only four of
the controlled trials included patients
with type 1 diabetes (57– 60). In addition,
there were 12 trials examining glycemic
parameters in healthy individuals. The re-
maining studies were 36 uncontrolled
prospective cohort trials and 2 case reports.

We present the available evidence for
26 of the substances with either one or
more controlled clinical trials in patients
with diabetes or impaired glucose toler-
ance. Methodological details and results
of these trials are summarized in Tables

1281

1282 Table 3—Controlled clinical trials of vitamin/mineral supplements for glycemic control* Review of herbs/vitamins in diabetes

Evidence Adverse Effects/

Herb/Supplement Reference Design Sample Intervention Control Outcomes Direction Jadad Events

Alpha-lipoic Acid Jacob S et al Blinding unclear; 4 74 Type 2; well-controlled Alpha-lipoic-acid 600 mg/d Placebo pill Increase glucose uptake; ϩ 1 No side effects
(1999) vs. 1200mg/d vs. 1800 trend decrease fasting
parallel groups on diet and/or OHA mg/d (Thioctacid, Asta insulin and improve
Medica, Germany); for 4 insulin sensitivity; no
wks change in FPG

Branched Chain AA Mourier A et al Open-label; 4 24 Type 2; on diet and/or Branched chain amino acid Placebo supplement No supplement effect on Ϫ 2 No side effects
(1997) supplement containing (tricalcic FBG, PPG, insulin,
parallel groups OHA leucine, isoleucine, valine phosphate and HgbA1C
(Paraphar Laboratories, stearate of
France) ϩ/Ϫ exercise magnesium)
training program; for 2 mos

Carnitine (Acetyl-L- Giancaterini A et Double-blind; 18 Type 2; on diet, OHA, Intravenous infusion acetyl-L- Saline infusion Increase glucose uptake, ϩϩ 4 Not reported
Carnitine) al (2000) Crossover; and/or insulin (switched camitine; 0.025mg/kg/min Saline infusion glucose storage;
Short-term to insulin during study) vs 0.1mg/kg/min; constant Saline infusion decrease insulin; no
metabolic trial infusion during change in glucose or
15 Type 2; on diet and euglycemic- lipid oxidation
Carnitine (L-Carnitine) Mingrone G et al Blinding unclear; OHA (switched to hyperinsulinemic clamp
(1999) Crossover; insulin during study) Increase glucose uptake, ϩϩ 1 Not reported
Short-term L-Carnitine; 0.28 ␮mol/kg glucose oxidation,
metabolic trial 9 Type 2 bw/min (Sigma Tau S.P.A., glucose storage,
Italy);simultaneous insulin sensitivity
Carnitine Capaldo B et al Blinding unclear; infusion with euglycemic
(1991) Crossover; hyperinsulinemic clamp Increase glucose uptake, ϩϩ 1 Not reported
Short-term insulin sensitivity
metabolic trial Carnitine; 1.7mmol/min;
constant intravenous
infusion with euglycemic
hyperinsulinemic clamp

Chromium Lee NA et al Double-blind; 30 Type 2; on diet, OHA, Chromium picolinate; 200␮g/ Placebo pill No change in FBG, Ϫ 4 No side effects
Chromium (1994) Crossover and/or insulin d (unspecified HgbA1C
preparation); for 2 mos
Chromium
Anderson R et al Double-blind; 3 180 Type 2; on diet, OHA, Chromium picolinate; 200␮g/ Matched placebo Decrease HgbA1C, ϩϩ 3 No side effects
Chromium (1997) parallel groups and/or TCM meds d vs. 1000␮g/d pill fasting and
Chromium (“Nutrition21,” San Diego, postprandial insulin
Chromium CA); for 8 wks (both doses);
decrease FBG and
DIABETES CARE, VOLUME 26, NUMBER 4, APRIL 2003 PPG (high dose)

Bahijiri SM et al Double-blind; 78 Type 2; on diet, OHA, Organic chromium (Brewer’s Torula yeast capsule Decrease FPG, PPG, ϩϩ 4 No side effects
(2000) Multiple and/or insulin yeast capsule 23.3␮g Cr/ fructosamine (both
crossover Cr supplement
day) vs. Inorganic types); no change in
BMI
chromium (CrCl3 capsule
200␮g Cr/day); for 8 wks

Uusitupa MIJ et al Double-blind; 2 26 elderly with impaired Chromium-rich yeast; 160␮g/ Identical placebo No change in FBG, PPG, Ϫ 3 Not reported
postprandial insulin,
(1992) parallel groups glucose tolerance d in 4 pellets (unspecified pellets HgbA1C, C-peptide,
BMI
preparation); for 6 mos

Anderson RA et al Double-blind; 8 impaired glucose Chromium Chloride; 200␮g/ Placebo tablet Decrease PPG, ϩϩ 2 Not reported

(1991) Crossover tolerance d (preparation postprandial insulin,
unspecified); for 4 wks glucagon

Cefalu WT et al Double-blind; 2 29 obese nondiabetic at risk Chromium picolinate; Placebo Increase insulin ϩ 2 No side effects
(1999) parallel groups sensitivity by
for Type 2 1000␮g/d (preparation FSIVGTT; no change

unspecified); for 8 mos

in FPG, PPG,
glycated Hgb,
fructosamine, weight;
trend decrease insulin

DIABETES CARE, VOLUME 26, NUMBER 4, APRIL 2003 Chromium Abraham AS et al Blinding unclear; 2 25 Type 2 with Chromium chloride; 250␮g/d Placebo supplement No change in FBG Ϫ 2 No side effects,
Chromium, Zinc no effect on
(1992) parallel groups atherosclerotic disease; in syrup (preparation in syrup liver/renal
Mg function tests,
Mg on diet and/or OHA unspecified); for 7–16 mos CBC,
Mg chemistries
Mg Anderson RA et al Double-blind; 4 110 Type 2; well-controlled Chromium pidolate 400␮g/d Placebo pill Decrease in plasma Ϫ
on diet, OHA, and/or vs. Zinc gluconate 30mg/d thiobarbituric acid 2 No side effects
Mg (2001) parallel groups insulin vs. ZnϩCr (Labcatal reactive substances
Mg Pharmaceutical, France); (TBARS); no change 3 No side effects
Mg, Vit C for 6 mos in FPG, HgbA1C,
Vanadium insulin, weight, BMI 3 Exanthem (1),
(all supplement gastrointestinal
Vanadium groups) pain (1)

de Lourdes LM et Double-blind; 3 128 Type 2, poorly Magnesium oxide; 20.7mmol/ Placebo pill Decrease fructosamine ϩ 2 Not reported
controlled (with d vs. 41.4 mmol/d Placebo pill (higher dose); no Ϫ
al (1988) parallel groups neuropathy and CAD) on elemental Mg; for 30 d Placebo pill change in FBG, 2 Not reported
diet and/or OHA HgbA1C, BMI
Eibl NL et al Double-blind; 2 Magnesium citrate; 30 mmol/ 2 No side effects
(1995) parallel groups 40 Type 2 with d (“Magnosolv granulate,” No change in HgbA1C,
hypomagnesemia; well- Asta Medica); for 3 mos FBG, PPG, insulin 1 Not reported
controlled on diet and
Paolisso G et al Double-blind; OHA Magnesium pidolate; 4.5g/d Decrease FBG; increase ϩϩ 3 No side effects
(1992) Crossover Mg, (“Mag2,” Lirca postprandial insulin, ϩ
12 nondiabetic (elderly Synthelabo, Italy); for 4 glucose uptake, N/A 5/6 transient
with insulin resistance) wks glucose oxidation; Ϫ gastrointestinal
unclear C-peptide ϩϩ discomfort; no
Paolisso G et al Double-blind; 9 Type 2, elderly, Magnesium pidolate; 4.5g/d Placebo effect on liver/
(1994) Crossover nonobese; on diet alone (“Mag2,” Lirca Synthelabo, Improve insulin kidney
Italy); for 4 wks sensitivity and function
glucose oxidation
deValk HW et al Blinding unclear; 2 50 Type 2; all on diet and Magnesium aspartate HCL; 15 Placebo during clamp; no N/A 7/7 transient
(1998) parallel groups insulin mmol/d (Verla-Pharm, Placebo pill change in FPG, C- gastrointestinal
Germany); for 3 mos No treatment peptide, glucagon, discomfort no
Paolisso G et al Open-label; 8 Type 2; on diet and OHA body weight effect liver/
(1999) Crossover (diet alone during study) Magnesium; 2gm/d (“Mag2,” kidney function
Lirca Synthelabo, Italy); for No change in FBG,
Erikksson J et al Double-blind; 29 Type 1, 27 Type 2; on 4 wks HgbA1C, urine
(1995) Crossover diet, OHA and/or insulin glucose
Magnesium (600 mg/day) vs.
Cohen N et al Ascorbic Acid (2g/day) Decrease FPG, increase
(1995) water soluble tablets; for postprandial insulin
90 d
Decrease FBG, HgbA1C Ϫ
(Vit C in Type 2
only); otherwise no
change

Non-randomized; 6 Type 2; diet and/or OHA Vanadyl sulfate hydrate; Placebo capsule Decrease FBG, HgbA1C, ϩϩ
Single-blind; 100mg/day (Spectrum hepatic glucose ϩϩ
Crossover Chemical, CA); for 3 wks production; increase
insulin-mediated
Halberstam M et Non-randomized; 7 Type 2 Vanadyl sulfate hydrate; Placebo capsule glucose uptake, Yeh and Associates
al (1996) Single-blind; 100mg/day (Spectrum insulin sensitivity;
Crossover Chemical, CA); for 3 wks trend decrease
fructosamine; no
1283 change PPG and C-
peptide

Decrease FBG, HgbA1C,
hepatic glucose
output; increase
insulin sensitivity; no
change in insulin

Review of herbs/vitamins in diabetesVanadiumBoden G et alNon-randomized; 8 Type 2; OHA and/orVanadyl sulfate; 100mg/d; for Placebo capsuleDecrease FBG, decrease ϩϩN/A 4/8 diarrhea; 1/8
(1996) 4 wks hepatic glucose nausea; 1/8
1284 Single-blind; insulin output during clamp flatulence

Crossover

Vit E Reaven PD et al Double-blind; 2 21 Type 2 men; on diet Vitamin E; 1600 IU/d dl- Placebo pill No change in FBG, PPG, Ϫ 4 No side effects

(1995) parallel groups and/or OHA alpha-tocopherol postprandial insulin,

(Hoffman-LaRoche); for 10 glycated Hgb,

wks glycated albumin,

glycated total plasma

proteins, fructos-

amine; decrease

susceptibility of LDL

to oxidation

Vit E Paolisso G et al Double-blind; 15 Type 2; well controlled Vitamin E; 900 mg/d dl- Sodium citrate Decrease HgbA1C, FPG, ϩϩ 3 No side effects

(1993) Crossover on diet and OHA alpha-tocopheryl acetate placebo PPG; no change in

(“Ephynal,” Roche, Italy); insulin, hepatic

for 4 mos glucose output,

glucose oxidation;

increase total body

glucose disposal and

non-oxidative glucose

metabolism

Vit E Gomez-Perez FJ et Double-blind; 53 DM (39 Type 2, 14 Type Vitamin E; 1200 mg/d (Searle Placebo capsule No change in FBG, Ϫ 3 Not reported

al (1996) Crossover 1); poorly controlled on de Mexico SA de CV): for 2 fructosamine,

diet, OHA and/or insulin mos HgbA1C

Vit E Paolisso G et al Double-blind; 25 Type 2; well controlled Vitamin E; 900 mg/d d-alpha- Placebo pill Decrease FPG, HgbA1C, ϩϩ 3 No side effects;

(1993) Crossover on diet and OHA tocopherol (“Ephynal,” PPG; no change in no effect on

Roche, Italy); for 3 mos insulin liver/renal

function tests

Vit E Ceriello A et al Single-blind; 3 30 “insulin-requiring DM”; Vitamin E; 1200mg/d vs. 600 Placebo Decrease Hgb A1C and ϩϩ 1 Not reported

(1991) parallel groups on diet and insulin mg/d (unspecified glycosylated protein

preparation); for 2 mos (dose related); no

change in FPG or

mean daily glucose

Vit E Jain SK et al Non-randomized; 35 Type 1 Vitamin E; 100 IU/d; for 3 Placebo capsule Decrease glycated Hgb; ϩ N/A Not reported

(1996) Double-blind; 2 mos no change FPG,

parallel groups insulin requirement

*All trials are randomized unless otherwise specified in the “Design” column. Ϫ, no outcome measures positive; ϩ, at least one outcome measure positive; ϩϩ, Ͼ50% of outcome measures positive. FBG, fasting
blood glucose; FSIVGTT, frequently sampled intravenous glucose tolerance test; PPG, postprandial glucose; OHA, oral hypoglycemic agent.

1–3 (Single herbs, Multiple herb combi-
nations, Vitamin and Mineral Supple-
ments). These tables include supplement
name, reference, study design, sample
population, intervention, control, out-
comes, direction of evidence, Jadad score,
and reported adverse effects. Other stud-
ies, including some RCTs that examined
glycemic parameters in healthy individu-
als, are not listed in the tables but are pre-
sented as supporting evidence when
applicable.

The most common outcomes mea-
sures encountered in these trials included
the following parameters of glycemic con-
trol: fasting and postprandial plasma glu-
cose, response to glucose tolerance tests
(GTTs), insulin and C-peptide levels, pro-
tein glycosylation (long-lived intracellu-
lar glycated hemoglobin and shorter-lived
plasma fructosamine), and clinical insulin
requirements. Many of the vitamin stud-
ies also examined measures of insulin sen-
sitivity, hepatic glucose production,
glucose oxidation, and glucose uptake.
Oftentimes, only a few of the above mea-
sures were studied in any particular trial.
A significant positive change in at least
one of these important parameters was re-
quired to categorize the trial as positive.

RESULTS

Single herbs/plant derivatives for
glycemic control
Table 1 presents the controlled clinical
trials of single herbs for glycemic control
in patients with diabetes. Of the single
herbs studied, the higher-quality RCTs
(with Jadad scores of 3 or greater) are
available for Coccinia indica, ginseng spe-
cies, Bauhinia forficata, and Myrcia uni-
flora. One RCT for Allium sativum is also of
adequate quality but was conducted in
nondiabetic individuals. Other herbs, Al-
lium cepa, Ocimum sanctum, Ficus carica,
Silibum marianum, Opuntia streptacantha,
and Trigonella foenum, have been studied
in poorer-quality RCTs. Gymnema sylves-
tre and Momordica charantia have been
studied in only nonrandomized con-
trolled trials.

Coccinia indica
Coccinia indica (ivy gourd) is a creeping
plant that grows wildly in many parts of
the India subcontinent, and is used to
treat “sugar urine” (madhumeha) in
Ayurveda, a traditional East Indian heal-
ing system. The mechanism of action of

DIABETES CARE, VOLUME 26, NUMBER 4, APRIL 2003

Yeh and Associates

Coccinia indica is not well understood, but (n ϭ 36 and n ϭ 24); both reported de- evidence for allium species in glycemic
the herb appears to have insulin-mimetic creases in fasting blood glucose and control. (Level I, C)
properties (61-63). HbA1c (68,69). Only one case of insomnia
was reported in these trials. Three other Ocimum sanctum
The one RCT of this herb (n ϭ 32), short-term metabolic trials in healthy vol- Ocimum sanctum (holy basil) is another
conducted in India, reported significant unteers also found decreases in postpran- commonly used herb in Ayurveda (relat-
changes in glycemic control following 6 dial glucose (66,70,71). All but one of the ed species include Ocimum album and
weeks’ use of powder from locally ob- clinical trials we examined were from the Ocimum basilicum). Studies in animal
tained crushed dried leaves in poorly con- same investigator group. The available ev- models suggest hypoglycemic effects
trolled or otherwise untreated patients idence for American ginseng in diabetes (77), although the mechanism of action
with type 2 diabetes (64). Another three- suggests a possible hypoglycemic effect; remains unknown. Postulated effects in-
arm controlled clinical trial (n ϭ 70) com- however, the trials are small and longer- clude enhanced ␤-cell function and insu-
pared 12 weeks’ use of dried herb pellets term studies are needed. (Level I, A) lin secretion. The one available controlled
made from fresh leaves with no treatment clinical trial of Ocimum sanctum (n ϭ 40)
and oral hypoglycemic agents (chloprop- Allium species: sativum and cepa showed positive effects on both fasting
amide) in patients with type 2 diabetes Allium sativum (garlic), a member of the and postprandial glucose in patients with
(61). The magnitude of change seen with lily family, is most commonly used world- type 2 diabetes using a local preparation
the herb was similar to that with a con- wide for flavorful cooking. Much of the of fresh leaf powder mixed in water for 4
ventional drug. Two other open-label clinical literature on garlic has focused on weeks (78). No adverse effects were re-
prospective trials offer supporting evi- its potential antioxidant activity and mi- ported. Further information is needed be-
dence of a hypoglycemic effect (62,63). crocirculatory effects (e.g., allicin and fore the efficacy of Ocimum sanctum in
No adverse events were reported in these ajoene for use in hypertension and hyper- diabetes can be fully assessed. (Level III,
trials. The preliminary evidence suggests lipidemia). Few studies have examined its C)
that the potential role for Coccinia indica in effects on insulin and glucose handling,
diabetes warrants further study. (U.S. although some attention has been given to Trigonella foenum graecum
Preventive Services Task Force Level I, allyl propyl disulfide, a volatile oil, and Trigonella foenum graecum (fenugreek) is a
American Diabetes Association Guide- S-allyl-cysteine sulfoxide, a sulfur con- legume extensively cultivated in India,
lines Level A) taining amino acid (39,72). Experiments North Africa, and the Mediterranean. It is
in animal models with alloxan-induced a common condiment used in Indian
Ginseng species diabetes have shown moderate reduc- cooking and commonly used herb in
Several different plant species are often tions in blood glucose; no effect is seen in Ayurveda. Defatted seeds of fenugreek,
referred to as ginseng. These include Chi- pancreatectomized animals (72,72). Al- which are rich in fiber, saponins, and pro-
nese or Korean ginseng (Panax ginseng), lium cepum (onion) also contains allyl pro- tein, have been described in early Greek
Siberian ginseng (Eleutherococcus sentico- pyl disulphide and has similar purported and Latin pharmacopoeias for hypergly-
sus), American ginseng (P. quiquefolius), hypoglycemic properties. Reported mecha- cemia. Although the seed portion is often
and Japanese ginseng (P. japonicus). Panax nisms of allium species include increased mentioned, other parts of the herb have
species (from the root panacea) are often secretion or slowed degradation of insu- also been investigated. Purported mecha-
touted for their “cure-all” adaptogenic lin, increased glutathione peroxidase ac- nisms include delay of gastric emptying,
properties, immune-stimulant effects, tivity, and improved liver glycogen storage slowing carbohydrate absorption, and
and their ability to increase stamina, con- (39,41). inhibition of glucose transport from the
centration, longevity, and overall well- fiber content, as well as increased eryth-
being (37). Preparations use the herb’s The highest quality RCT of Allium sa- rocyte insulin receptors and modulation
root; some sources report greater efficacy tivum in humans was actually designed to of peripheral glucose utilization. Many
with roots that are greater than 3 years examine thrombocyte aggregation in studies in alloxan-rat models have shown
old. Principal components are believed to nondiabetic individuals (n ϭ 60). How- modulated exocrine pancreatic secretion
be the triterpenoid saponin glycosides ever, the investigators found significant (79).
(ginsenosides or panaxosides). Hypogly- decreases in fasting serum glucose (74).
cemic effects have been shown in strepto- The only available trial of garlic in patients There are several trials available for
zotocin rat models (65). Reported mech- with type 2 diabetes (n ϭ 33) did not find fenugreek in type 2 diabetes; however,
anisms of action include decreased rate of consistent glucose or insulin responses af- most are noncontrolled (158). Of the
carbohydrate absorption into the portal ter 1 month of supplementation (75). The available RCTs, they are generally poorer-
hepatic circulation, increased glucose only clinical trial available for Allium cepa quality studies with small numbers (n ϭ
transport and uptake mediated by nitric is a small RCT of allyl propyl disulphide 5–15) and from a single investigator
oxide, increased glycogen storage, and extract capsules from onion in nondia- group. Nonetheless, these trials, includ-
modulation of insulin secretion (39). betic volunteers (n ϭ 6); investigators ing a single trial in type 1 diabetes, have
showed an acute decrease in fasting blood reported improved glycemic control us-
Most clinical trials we found utilized glucose and increase in insulin, support- ing seed powder incorporated into un-
American ginseng, with many examining ing an insulin-mediated effect (76). No leavened bread (59,80). Another trial in
the herb’s short-term effects on patients adverse events were reported in these healthy volunteers (n ϭ 38) incorporated
with type 2 diabetes after a standard oral trials. The limited data provide conflicting several short-term randomized crossover
GTT (66,67). Two longer-term trials ad- experiments administering different
ministered American ginseng for 8 weeks

DIABETES CARE, VOLUME 26, NUMBER 4, APRIL 2003 1285

Review of herbs/vitamins in diabetes

preparations of fenugreek before oral decrease in postprandial glucose and in- tes (n ϭ 60) using a commercially
GTT. In these series of trials, whole raw sulin requirements, but no change in fast- available preparation (“Legalon” 600 mg/
seeds, extracted seed powder, gum isolate ing glucose when compared with the day; IBI Lorenzini, Milan, Italy) for 12
of seeds, and cooked whole seeds seemed control commercial tea (60). No effect months, with significant improvements
to decrease postprandial glucose levels, was seen in C-peptide levels, thereby sup- in glycemic control when compared with
whereas degummed seeds and cooked porting a non–insulin-mediated effect. no treatment (92). No adverse effects
leaves did not (79). Other open-label pro- No adverse effects were reported. Clearly, were reported. Further information and
spective cohort studies have followed pa- more information is needed before the ef- higher quality clinical trials are needed to
tients on fenugreek for up to 6 months ficacy of Ficus carica can be properly as- further investigate milk thistle in glycemic
with reported benefits in glycemic control sessed. (Level III, C) control. (Level III, C)
(79,81– 84). No adverse effects were re-
ported in these trials. There is some pre- Opuntia streptacantha Gymnema sylvestre
liminary evidence for the efficacy of Opuntia streptacantha (nopal) or the Gymnema sylvestre is another commonly
fenugreek that suggests further studies prickly pear cactus can be found in arid used herb in Ayurveda. The plant is a
may be warranted. (Level II-2, C) regions throughout the Western hemi- woody climber that grows in tropical for-
sphere, including the southwestern U.S., ests of central and southern India. Ac-
Bauhinia forficata and Myrcia and is commonly used for glucose control cording to common folklore, chewing the
uniflora by those of Mexican descent. It has a high- leaves causes a loss of sweet taste, hence
Indigenous to rainforests and tropical ar- soluble fiber and pectin content, which the popular Hindi name of the plant “gur-
eas of South America, Bauhinia forficata may affect intestinal glucose uptake, par- mar,” meaning “destroyer of sugar.” Early
has been used in traditional treatment of tially accounting for its hypoglycemic ac- animal studies reported blood glucose–
diabetes in that area. In Brazilian herbal tions (65). Animal models have reported lowering effects in animals with residual
medicine, Bauhinia species have been re- decreases in postprandial glucose and pancreatic function, but no effect in total
ferred to as “vegetable insulin.” Another HbA1c with synergistic effects with insu- pancreatectomized animals. Studies of an
commonly used South American herb is lin. Studies in pancreatectomized animals ethanol leaf extract, GS4, in diabetic rat
Myrcia uniflora. As part of a national effort report that hypoglycemic activity is not and rabbit models have reported regener-
to identify potential plant species useful dependent on the presence of insulin ation of islets of Langerhans, decreases in
in glucose control, two small crossover (89). Most human studies of nopal have blood glucose, and increases of serum in-
studies (n ϭ 16 and n ϭ 18) by one in- been published in Spanish and, thus, are sulin (58). Mechanism of action is un-
vestigator administered each of these not included in this review. We found known; postulated theories include an
herbs as tea infusions to separate groups only two controlled short-term metabolic increase in glucose uptake and utilization,
of patients three times daily for 8 weeks. trials (n ϭ 14 and n ϭ 32) published in increase in insulin release through cell
No significant differences in glucose or the English language, both by the same permeability, increase in ␤-cell number,
HbA1c were detected between study herb investigator (90,91). These reported im- and stimulation of ␤-cell function (39,93).
infusion and a placebo tea using Imperata provements in patients with type 2 diabe-
brasiliensis. No adverse effects were re- tes with decreased fasting glucose and Two nonrandomized controlled clin-
ported (85). This limited preliminary evi- decreased insulin levels, suggesting en- ical trials are available, both from the
dence does not support the hypoglycemic hanced insulin sensitivity. No side effects same investigator group. Groups of pa-
effect of these two plant species. (Level I, were reported in these trials. The limited tients with type 1 diabetes (n ϭ 64) and
American Diabetes Association level not data suggests a possible hypoglycemic ef- type 2 diabetes (n ϭ 47) showed im-
applicable if no studies show benefit) fect of nopal; however, longer-term clini- proved glycemic control with chronic ad-
cal trials are needed. (Level III, C) junctive use of GS4 extract compared
Ficus carica with those who received conventional
Ficus carica (fig leaf) is a popular plant Silibum marianum treatment alone (58,94). The evidence for
used for patients with diabetes in Spain Silibum marianum (milk thistle), a mem- the beneficial effect of Gymnema sylvestre
and other areas in Southwestern Europe. ber of the aster family, has been primarily in diabetes is suggestive, although incon-
Its active component is unknown. Several studied for its purported effects on alco- clusive given the limited data. (Level II-1,
studies in animal models with diabetes holic and viral hepatitis, rather than for C)
have shown both short- and long-term glycemic control. However, silymarin is
hypoglycemic effects, although human rich in flavonoids, potent antioxidants, Momordica charantia
trials are lacking. Potential hypolipdemic and some have postulated a potential ben- Momordica charantia is a vegetable indig-
effects in diabetic rats have also been efit for those who have insulin resistance enous to tropical areas, including India,
shown (86 – 88). Its mechanism for glu- secondary to hepatic damage (39). Mech- Asia, South America, and Africa, also
cose effect is unknown; however, some anisms are based on the herb’s antioxi- known as balsam pear, karela (karolla),
studies suggest facilitation of glucose up- dant activity and effects on hepatocyte and bitter melon. Reported preparations
take peripherally. The one available clin- stabilization with decreased glutathione of the herb range from injectable extracts
ical trial is a small crossover study of fig oxidation, as well as on restoration of nor- to fruit juice to fried melon bits (39,95–
leaf tea for 4 weeks in patients with type 1 mal malondialdehyde concentrations. 97). Active components are thought to be
diabetes (n ϭ 10); investigators showed a charantin, vicine, and polypeptide-p (an
The one available clinical trial exam- unidentified insulin-like protein similar
ined cirrhotic patients with type 2 diabe- to bovine insulin). Theoretical mecha-

1286 DIABETES CARE, VOLUME 26, NUMBER 4, APRIL 2003

Yeh and Associates

nisms include increased insulin secretion, Multiple herb combinations for reported with this formulation. The avail-
tissue glucose uptake, liver muscle glyco- glycemic control able studies suggest that some TCM for-
gen synthesis, glucose oxidation, and de- Table 2 presents the controlled clinical mulations, but not others, may have
creased hepatic gluconeogenesis. Studies trials of multiple herb combinations for beneficial effects. However, the data are
in alloxan-induced diabetic rabbits have glycemic control in patients with diabetes. certainly limited and no formula has been
suggested hypoglycemic effects (98). studied in more than one trial. (Level I, C)
Combination formulas in TCM
Two controlled short-term metabolic TCM encompasses a system of healing Combination formulas in Native
trials in patients with type 2 diabetes (n ϭ that has origins over 2,000 years old. It American medicine
18 and n ϭ 9) have reported acute effects emphasizes the importance of a balanced Native American medicine refers to the
on blood glucose with Momordica charan- and harmonious flow of “qi,” or “life healing practices from the people indige-
tia fruit juice, as well as subcutaneous force,” and employs diverse modalities nous to North America; the approach
vegetable insulin extract (95,99). Two such as acupuncture, massage, qigong, combines awareness of mind, body, and
other small, uncontrolled open-label tri- and an individualized approach to herbal spirit and ritualistic observances with
als also reported positive effects on glyce- medicine (20). We found few trials of practices of herbalism. One clinical trial
mic control after longer-term use (7–11 TCM in the English language; most have (n ϭ 40) specifically examined an herbal
weeks) (96,97). No adverse effects were been published in Chinese and were un- tea preparation containing Populus tremu-
reported in these trials. Some, albeit lim- available for this review. loides (trembling aspen) and Heracleum
ited, data suggest a potential effect of Mo- lanatum (cow parsnip) prescribed by an
mordica charantia in diabetes. However, One controlled clinical trial of a mul- Alexis band Sioux healer (117). Investiga-
further information in RCTs is needed. tiple herb combination examined a specific tors reported no glycemic benefit over a
(Level III, C) formulation containing Coptis chinensis, control tea containing mint and fennel
Astragalus membranaceus, and Lonicera ja- seed. Little is known scientifically about
Aloe vera ponica. Among a host of other plants used this formula, and it has not been studied
Aloe vera is the most well-known species in TCM for the treatment of diabetes, elsewhere. The limited evidence for this
of aloe, a desert plant resembling the cac- these plants were selected for study by the Native American herb preparation does
tus in the Liliaceae family. It is popularly Chinese Academy of Medical Science not support its use in glycemic control.
used to treat burns and promote wound based on experiential reports of efficacy (Level I, American Diabetes Association
healing. The dried sap of the Aloe vera is a and safety. Mechanisms of action are not level not applicable if studies show no
traditional remedy for diabetes in the Ara- well reported, but may include decreasing benefit)
bian peninsula (33), although aloe gel is digestive carbohydrate absorption. This
preferred over the sap as the latter con- formula is not thought to influence action Combination formulas in Tibetan
tains the laxative anthraquinone (100). of insulin. Using a 2ϫ2 factorial design medicine
Aloe gel, obtained from the inner portion (n ϭ 216) with TCM verum pill or pla- Tibetan medicine is a traditional system of
of the leaves, contains glucomannan, a cebo and glibenclamide verum pill or pla- healing that has influences from China,
hydrosoluble fiber which may in part ac- cebo, investigators reported that the two Persia, India, and Greece, incorporating
count for its hypoglycemic effects (39). treatments together were more efficacious concepts from Ayurveda as well as psy-
Reports in animal models have been in- than either alone (114). Of 216 patients, chological, philosophical, and spiritual
consistent (100 –103). Two nonrandom- there was one report of diarrhea and one aspects of Buddhism. Herbalism, espe-
ized clinical trials (n ϭ 40 and n ϭ 76) are report of dry mouth. Also, one case of cially from the Himalayas, plays an im-
available from the same investigator hypoglycemia occurred in the combined portant role. Although of poorer quality,
group that reported improved fasting treatment group. one large RCT (n ϭ 200) was available
blood glucose with 6 weeks of juice made that examined individualized Tibetan
from aloe gel (100,104). Case reports of A much smaller trial (n ϭ 12) of lower herb prescription based on age, sex, per-
five type 2 diabetic individuals reported quality examined another TCM prepara- sonality, pulse, and urine characteristics
decreases in fasting blood glucose as well tion, Xiaoke tea. Little is written about this in traditional diagnosis (118). Individual
as HbA1c (101). No adverse effects were formulation in English literature. It ap- plant species and postulated mechanisms
reported in these trials. The preliminary pears not to affect insulin concentrations were not reported. At 6 months, the study
data suggest a potential effect of Aloe vera and was ineffective in rats that lack en- suffered a large number of dropouts
in glycemic control; however, further in- dogenous insulin. The trial did not report (44%); however, investigators analyzed
formation is needed. (Level II-1, C) details about the constituents of the treat- data by intention-to-treat, and improve-
ment tea, and investigators reported no ments were nevertheless reported in fast-
Other herbs that have been studied difference in glycemic parameters as com- ing plasma glucose, postprandial glucose,
solely in uncontrolled trials include ber- pared with an “ordinary” tea infusion and HbA1c values. No adverse effects
berine (105), Cinnamomym tamala (106), (115). Another controlled clinical trial were noted. These limited data are incon-
curry (107), Eugenia jambolana (108), (n ϭ 148) examined a formulation called clusive regarding use of individual
gingko (109), Phyllanthus amarus (110), Semen Persical Decoction for Purgation Tibetan herb prescriptions in type 2 dia-
Pterocarpus marsupium (111), Solanum with Addition (SPDPA), a combination of betes. (Level II-2, C)
torvum (112), and Vinca rosea (113). eight different herbs and reported de-
creases in fasting blood glucose not signif- We identified six other specific com-
icantly different from changes seen with bination herb formulations that have
glyburide (116). No adverse effects were

DIABETES CARE, VOLUME 26, NUMBER 4, APRIL 2003 1287



Yeh and Associates

Vanadium trial showed no changes in fasting blood mostly for supportive evidence from
Vanadium has been described as either a glucose (153). Another noncontrolled RCTs with methodological flaws or un-
nonessential nutrient or a nutrient that is trial offers supportive evidence for a controlled studies, or conflicting evidence
required only in minute quantities, as no change in insulin sensitivity (152). The with weight supporting the recommenda-
physiological role of the trace element has available data are limited and suggest that tion (online appendix B). Those supple-
yet to be found (35,149). Human defi- further elucidation of ␣-lipoic acids ac- ments that earned an A rating include
ciency has not been documented. There tions is needed. (Level II-3, C) Coccinia indica, American Ginseng, and L-
are no accurate assays in clinical settings, carnitine, with supportive evidence from
and there is no recommended daily allow- DISCUSSION — A total of 108 hu- at least one adequate RCT. However, ac-
ance. Vanadium exists in several valence man trials of herbs and vitamin/mineral cording to the criteria described by
forms, with vanadyl (ϩ5) sulfate and so- supplements for glycemic control were Weiger et al. (56), no herb or supplement
dium metavanadate (ϩ4) being the most obtained. Most trials examined supple- has sufficient evidence to actively recom-
common supplement forms. Its mecha- ments as an adjunct to conventional treat- mend or discourage its use among pa-
nism of action in glycemic control is ment with diet and/or medication. Of the tients with diabetes. That is, evidence
thought to be primarily insulin-mimetic available trials, 58 were controlled (42 regarding efficacy is inconclusive or not
with upregulation of insulin receptors. In RCTs) and conducted specifically in indi- rigorous enough to meet the outlined re-
animal models, it has been shown to fa- viduals with diabetes or impaired glucose quirements of efficacy, yet the herb or
cilitate glucose uptake and metabolism tolerance. Among these controlled trials, supplement appears to be generally safe.
and to enhance insulin sensitivity. Clini- statistically significant treatment effects Physicians should thus keep an open
cally, it may enhance glucose oxidation were reported in 88% (23 of 26) of those mind in advising patients who might al-
and glycogen synthesis, and it may mod- examining single herbs, 60% (3 of 5) of ready be using these supplements.
ulate hepatic glucose output (35). Three those examining combination herbs, and
very small controlled clinical trials (n ϭ 67% (18 of 27) of those examining vita- The American Diabetes Association
6 – 8) have reported decreased fasting min and mineral supplements. However, and the American Dietetic Association do
blood glucose (138 –140); two of these many trials were of poor quality. More not have specific recommendations for
trials also reported significant changes in than half of the RCTs (24 of 42, 57%) the use of herb or vitamin/mineral supple-
HbA1c and insulin sensitivity (138,139). scored 2 or less on the Jadad scale. (No ments in people with diabetes. Broad rec-
Two noncontrolled open-label studies, RCT achieved a score of 5.) Thirteen trials ommendations for the general public are
also with small sample sizes, nonetheless had sample sizes of 10 or fewer patients. that healthy people at low risk for nutri-
offer supporting evidence (150,151). In addition, there were generally few trials tional deficiencies meet their require-
Goldfine et al. (151) included type 1 dia- per supplement, making it difficult to ments with natural food sources. Those at
betic patients (n ϭ 5) who decreased their draw definitive conclusions regarding ef- increased risk for deficiencies, such as the
insulin requirements after 2 weeks of ficacy. Nevertheless, no major safety con- elderly, strict vegetarians, those following
treatment. Gastrointestinal discomfort, cerns were reported in these trials. Few very low-calorie diets, and other special
including diarrhea, nausea, and flatu- mild adverse effects, mainly gastrointesti- populations, may benefit from multivita-
lence, was reported by a large proportion nal irritation, were reported for ginseng, min supplements (35).
of patients in all the vanadium trials. Or- Native American herb tea, TCM pill, mag-
ganically chelated compounds, however, nesium, and vanadium (see Tables). For Despite the lack of formal recommen-
are thought to cause less gastrointestinal the following supplements, Ͼ50% of dations, the American Diabetes Associa-
irritation than vanadium salts (149). The controlled clinical trials (at least two tri- tion has acknowledged patient interest
evidence for efficacy of vanadium in glu- als) suggested efficacy: Coccinia indica, and use of CAM supplements for diabetes.
cose control is suggestive, but as yet no Trigonella foenum, American ginseng, no- In A Step-by-Step Approach to Complemen-
RCTs are available. (Level II-1, C). pal, Gymnema sylvestre, Aloe vera, Mo- tary Therapies and Guidelines for Using Vi-
mordica charantia, chromium, and tamin, Mineral, and Herbal Supplements
␣-Lipoic acid vanadium. Of these, the best evidence is (154,155), safety is the main theme. Prac-
Also known as thioctic acid, a disulfide available for Coccinia indica and American tical information for patients on choosing
compound synthesized in the liver, ␣-li- ginseng. Supplements that appear effec- supplements is outlined (e.g., looking for
poic acid is a potent lipophilic antioxi- tive but have only been studied in non- products with recognized symbols of
dant. It is a cofactor in many multienzyme randomized trials include Gymnema quality: USP, NF, TruLabel, Consumer-
complexes and may also play a role in sylvestre, Aloe vera, and vanadium. Sup- Labs, etc.; looking for products with an
glucose oxidation (152). Experimental in plements that appear to be effective in expiration date; avoiding foreign prod-
vitro data have shown possible effects in short-term metabolic trials include Mo- ucts unless quality is known; and avoid-
enhancing glucose uptake in muscle and mordica, nopal, and L-carnitine. ing companies that make sensational
preventing glucose-induced protein claims or have misleading labels, etc). The
modifications. One multiple-dosage con- Guidelines for clinicians American Diabetes Association also warns
trolled trial is available in patients with In assessing the quality of the evidence, against combining supplements and pre-
type 2 diabetes (n ϭ 74), and it reported we employed the American Diabetes As- scription drugs without the physician’s
positive effects on glucose uptake and in- sociation criteria for clinical guidelines knowledge and against stopping pre-
sulin sensitivity with 600 –1,800 mg/day (55). The evidence for the majority of scribed medication without the physician’s
␣-lipoic acid for 4 weeks; however, the supplements earned a C level rating, knowledge. They advise discontinuing
supplements before medical procedures

DIABETES CARE, VOLUME 26, NUMBER 4, APRIL 2003 1289

Review of herbs/vitamins in diabetes

(e.g., surgeries or anesthesia) and in the dardize supplements, there is a general nisms of action so that applicability to
event of an adverse effect. lack of consistency across the market. type 1 or type 2 diabetes can be clarified.
With vitamin and mineral supplements,
Although the trials contained in this these issues are less relevant. CONCLUSIONS — As interest in the
review reported very few adverse effects, potential benefit of herbs and supple-
other sources mentioned potential or In addition, the development of ments for diabetes grows, it will become
theoretical effects for six supplements. proper supplement regulation and safety increasingly important to monitor the
Theoretical cross-allergenicity was men- codes has been slow. Currently, all dietary progress of the clinical literature and to
tioned with silymarin as a member of the supplements (including herbal products) communicate these findings to patients.
aster family (daisy) and Trigonella as a are regulated under the Dietary Supple- Based on this review, there is insufficient
member of the leguminosae family (pea- ment Health and Education Act of 1994 evidence to actively recommend or dis-
nuts), although no actual cases have been (DSHEA), which specifically differenti- courage use of any particular supplement,
reported. The most important potential ates supplements from drugs. Conse- although most appeared to be generally
drug-herb interaction was that of garlic or quently, DSHEA does not require the safe. Preliminary evidence of several
Trigonella with warfarin, as both herbs extensive premarket approval that the herbs and supplements suggest that fur-
may have limited anticoagulant proper- Food and Drug Adminstration requires ther RCTs may be warranted. The seven
ties. Momordica may increase risk of po- for a prescription drug, and although it most promising supplements include
tassium depletion, so caution might be calls for “good manufacturing practices Coccinia indica, American ginseng, Mo-
taken with those on laxatives or diuretics. [GMP],” the burden of proof that a sup- mordica charantia, nopal, L-carnitine,
Ginseng used in conjunction with mono- plement is unsafe lies with the govern- Gymnema sylvestre, Aloe vera, and vana-
amine oxidase inhibitors, phenelzine, or ment, leaving manufacturers to operate dium. Until more definitive studies help
stimulants may cause an enhanced eu- unchecked. This has contributed to skep- to clarify our questions, clinicians should
phoric effect. Other adverse effects have ticism among clinicians, and makes it es- remain cautious, yet open-minded, re-
been reported with Panax ginseng (Asian) pecially difficult for physicians to garding adjunctive use of these supple-
(e.g., hypertension, hypotension, mastal- responsibly recommend supplements to ments. They should be guided not only by
gia, vaginal bleed, and insomnia), al- patients. In the absence of external regu- sound clinical judgement, but also by pa-
though the literature on diabetes has lation, the industry has taken steps to po- tients’ preferences, needs, and values. As
largely involved Panax quiquefolius lice itself. For example, the National we further our understanding of herbs
(American). Rare topical reactions have Nutritional Foods Association (NNFA), and dietary supplements, we might begin
been reported with nopal, garlic, and ␣-li- representing about one-third to one-half to develop a framework for a medical sys-
poic acid. Of note, one case of hypoglyce- of retailers and manufacturers of natural tem capable of incorporating those com-
mic coma has been reported with products in the U.S., has encouraged the plementary therapies proven to be
overdosage of Momordica charantia adoption of strict, self-imposed GMP beneficial.
(36,37,39, www.naturaldatabase.com). standards, as well as initiatives such as the
TruLabel program (in which products are Addendum — Since our review of this topic,
Clinical research of CAM subjected to random laboratory testing by the report of a large multicenter trial (n ϭ
supplements in diabetes independent third-party auditors to ver- 3,654), which examined the effects of vitamin
Currently, there is not yet sufficient eval- ify contents) (42). E with and without ramipril in high-risk pa-
uation of herbs, vitamins, and mineral tients with diabetes, has been published. Al-
supplements for glucose control in diabe- Research of vitamin and mineral sup- though this study was primarily concerned
tes. Aside from relatively poor study plements has also been hindered by a lack with cardiovascular events and mortality, it
methodological quality, this area of sup- of accurate and meaningful assays that de- does report that there were no differences in
plement research has been fraught with tect functional micronutrient deficiencies. change of HbA1c between groups (157).
several complications. In the case of chromium, for example, it is
postulated that supplementation of tar- Acknowledgments — The authors thank Dr.
First, the multiple constituent nature geted individuals might be more benefi- Alan Moses and Karen Chalmers for their
of botanical products has made standard- cial. Some speculate that positive results thoughtful review of the manuscript.
ization a challenging task. Proponents of seen in large studies in diabetic patients in
herbal remedies caution that in standard- China may be due to the population’s rel- References
izing to one constituent, resulting extracts ative chromium deficiency. However, 1. National Diabetes Fact Sheet: National Es-
may have lost a proportion of benefit as without reliable assays, these theories timates and General Information on Diabe-
compared with the whole plant (156). have remained difficult to test (144). tes. Atlanta, Georgia, Center for Disease
Precise considerations of purity, chemical Control and Prevention, 1998
composition, and potency of derivatives Finally, the existing literature in this 2. Eisenberg DM, Davis RB, Ettner SL, Ap-
may be grossly influenced by the age of area includes a considerable amount of pel S, Wilkey S, Van Rompay M, Kessler
the plant (especially of roots), the source study population heterogeneity. Future RC: Trends in alternative medicine use
location, the season of harvest, the research may need to more precisely de- in the United States, 1990 –1997: results
method of drying and crude preparation, fine targeted diabetic populations with re- of a follow-up national survey. JAMA
etc. In the literature we examined, several gard to disease classification, severity, 280:1569 –1575, 1998
herb studies used “homemade” or other- optimal adjunctive interventions, and 3. Payne C: Complementary and integra-
wise unspecified preparations. Although perhaps nutrient deficiencies. It will also
individual companies have begun to stan- be important to further elucidate mecha-

1290 DIABETES CARE, VOLUME 26, NUMBER 4, APRIL 2003



Review of herbs/vitamins in diabetes

necessary? Control Clin Trials 17:1–12, 217–226, 2002 subjects. Clin Chim Acta 60:121–123,
1996 66. Vuksan V, Sievenpiper JL, Koo VYY, 1975
53. Moher D, Jadad AR, Nichol G, Penman 77. Chattopadhyay RR: Hypoglycemic effect
M, Tugwell P, Walsh S: Assessing the Francis T, Beljan-Zdravkovic U, Xu Z, of Ocimum sanctum leaf extract in normal
quality of randomized controlled trials: Vidgen E: American ginseng (Panax and streptozotocin diabetic rats. Ind J
an annotated bibliography of scales and quinquefolius L) reduces postprandial Exp Biol 31:891– 893, 1993
checklists. Control Clin Trials 16:62–73, glycemia in non-diabetic subjects and 78. Agrawal P, Rai V, Singh RB, Azad Khan
1995 subjects with type 2 diabetes mellitus. AK, Akhtar S, Mahtab H: Randomized
54. US Preventive Services Task Force: Arch Intern Med 160:1009 –1013, 2000 placebo-controlled single-blind trial of
Guide to Clinical Preventive Services: An 67. Vuksan V, Stavro MP, Sievenpiper JL, holy basil leaves in patients with non-
Assessment of the Effectiveness of 169 Inter- Beljan-Zdravkovic U, Leiter LA, Josse insulin-dependent diabetes mellitus:
ventions. Baltimore, MD, Williams and RG, Xu Z: Similar postprandial glycemic Coccinia indica in the treatment of pa-
Wilkins, 1989 reductions with escalation of dose and tients with diabetes mellitus. Int J Clin
55. American Diabetes Association: Stan- administration time of American gin- Pharmacol Ther 34:406 – 409, 1996
dards of medical care for patients with seng in type 2 diabetes. Diabetes Care 79. Sharma RD: Effect of fenugreek seeds
diabetes mellitus (Position Statement). 23:1221–1226, 2000 and leaves on blood glucose and serum
Diabetes Care 25 (Suppl. 1):S33–S49, 68. Sotaniemi EA, Haapakoski E, Rautio A: insulin responses in human subjects.
2002 Ginseng therapy in non-insulin de- Nutr Res 6:1353–1364, 1986
56. Weiger WA, Smith M, Boon H, Richard- pendent diabetic patients: effects on 80. Sharma RD, Raghuram TC: Hypoglyce-
son MA, Kaptchuk TJ, Eisenberg DM: psychophysical performance, glucose mic effect of fenugreek seeds in non-in-
Advising patients with cancer who seek homeostasis, serum lipids, serum amino- sulin dependent diabetic subjects. Nutr
complementary/alternative/integrative terminalpropeptide concentration, and Res 10:731–739, 1990
medical therapies: an evidence-based body weight. Diabetes Care 18:1373– 81. Kuppu Rajan K, Srivatsa A, Krish-
approach. Ann Intern Med 137:889 – 1375, 1995 naswami CV, Bagavathiammal KC,
903, 2002 69. Vuksan V, Sievenpiper JL, Xu Z, Wong Janaki I, Vijayakumar G, Chellamariap-
57. Jain SK, McVie R, Jaramillo JJ, Palmer M, EY, Jenkins AL, Beljan-Zdravkovic U, pan M, Asha Bai PV, Buchi Babu O: Hy-
Smith T: Effect of a modest vitamin E Leiter LA, Josse RG, Stavro MP: Konjac- poglycemic and hypotriglyceridemic
supplementation on blood glycated he- mannan and american ginseng: emerg- effects of Methika churna (fenugreek).
moglobin and triglyceride levels and red ing alternative therapies for type 2 Antiseptic 95:78 –79, 1998
cell indices in type I diabetic patients. diabetes mellitus. J Am Coll Nutr 20: 82. Sharma RD, Sarkar A, Hazra DK, Mishra
J Am Coll Nutr 15:458 – 461, 1996 370S–380S, 2001 B, Singh JB, Sharma SK, Maheshwari BB,
58. Shanmugasundaram ERB, Rajeswari G, 70. Vuksan V, Sievenpiper JL, Wong J, Xu Z, Maheshwari PK: Use of fenugreek seed
Baskaran K, Kumar BRR, Shanmu- Beljan-Zdravkovic U, Arnason JT, Assin- powder in the management of non-insu-
gasundaram KR, Ahmath BK: Use of ewe V, Stavro MP, Jenkins AL, Leiter LA, lin dependent diabetes mellitus. Nutr Res
Gymnema sylvestre leaf extract in the Francis T: American ginseng (Panax 16:1331–1339, 1996
control of blood glucose in insulin-de- quinquefolius L) attenuates postpranidal 83. Sharma RD, Sarkar A, Harza DK, Misra
pendent diabetes mellitus. J Ethnophar- glycemia in a time-dependent but not B, Singh JB, Maheshwari BB, Sharma SK:
macology 30:281–294, 1990 dose-dependent manner in healthy indi- Hypolipidemic effect of fenugreek seeds:
59. Sharma RD, Raghuram TC, Rao NS: Ef- viduals. Am J Clin Nutr 73:753–758, a chronic study in non-insulin depen-
fect of fenugreek seeds on blood glucose 2001 dent diabetic patients. Phytother Res 10:
and serum lipids in type I diabetes. Eur 71. Vuksan V, Stavro MP, Sievenpiper JL, 332–334, 1996
J Clin Nutr 44:301–306, 1990 Koo VY, Wong E, Beljan-Zdravkovic U, 84. Sharma RD, Sarkhar DK, Hazra, Misra B,
60. Serraclara A, Hawkins F, Perez C, Francis T, Jenkins AL, Leiter LA, Josse Singh JB, Maheshwari BB: Toxicological
Dominguez E, Campillo JE, Torres MD: RG, Xu Z: American ginseng improves evaluation of fenugreek seeds: a long
Hypoglycemic action of an oral fig-leaf glycemia in individuals with normal glu- term feeding experiment in diabetic pa-
decoction in type-I diabetic patients. Di- cose tolerance: effect of dose and time tients. Phytother Res 10:519 –520, 1996
abetes Res Clin Pract 39:19 –22, 1998 escalation. J Am Coll Nutr 19:738 –744, 85. Russo EMK, Reichelt AAJ, De-Sa JR,
61. Kamble SM, Jyotishi GS, Kamlakar PL, 2000 Furlanetto RP, Moises RC, Kasamatsu
Vaidya SM: Efficacy of Coccinia indica 72. Sheela CG, Augusti KT: Antidiabetic ef- TS, Chacra AR: Clinical trial of Mycia
W.& A in diabetes mellitus. J Res fects of S-allyl cysteine sulphoxide iso- uniflora and Bauhinia forficata leaf ex-
Ayurveda Siddha. XVII:77– 84, 1996 lated from garlic Allium sativum Linn. tracts in normal and diabetic patients.
62. Kamble SM, Kamlakar PL, Vaidya S, Indian J Exp Biol 30:523–526, 1992 Braz J Med Biol Res 23:11–20, 1990
Bambole VD: Influence of Coccinia indica 73. Jain RC, Vyas CR, Mahatma OP: Hypo- 86. Campillo JE, Perez C, Ramiro JM, Torres
on certain enzymes in glycolytic and li- glycemic action of onion and garlic. Lan- MD: Hypoglycemic activity of an aque-
polytic pathway in human diabetes. In- cet 2:1491, 1973 ous extract from Ficus carica in strepto-
dian J Med Sci 52:143–146, 1998 74. Keisewetter H, Jung F, Pindur G, Jung zotocin diabetic rats (Abstract). Diabe-
63. Kuppurajan K, Seshadri C, Revathi R, EM, Mrowietz C, Wenzel E: Effect of gar- tologia 34 (Suppl.2):A-181, 1991
Venkataraghavah S: Hypoglycaemic ef- lic on thrombocyte aggregation, micro- 87. Torres MD, Dominguez E, Romero A,
fect of Coccinia indica in diabetes melli- circulation, and other risk factors. Int Campillo JE, Perez C: Hypoglycemic and
tus. Nagarjun 29:1– 4, 1986 J Clin Pharm Ther Tox 29:151–155, 1991 hypolipidemic activity of an aqueous ex-
64. Azad Khan AK, Akhtar S, Mahtab H: 75. Sitprija S, Plengvidhya C, Kangkaya V, tract from Ficus carica in streptozotocin
Coccinia indica in the treatment of pa- Bhuvapanich S, Tunkayoon M: Garlic diabetic rats (Abstract). Diabetologia 36
tients with diabetes mellitus. Bangladesh and diabetes mellitus phase II clinical (Suppl. 1):A-181, 1993
Med Res Counc Bull 5:60 – 66, 1979 trial. J Med Assoc Thai 70:223–227, 1987 88. Perez C, Dominguez E, Ramiro A,
65. Shapiro K, Gong WC: Natural products 76. Augusti KT, Benaim ME: Effect of essen- Campillo J, Torres MD: A study on the
used for diabetes. J Am Pharm Assoc 42: tial oil of onion on blood glucose, free glycemic balance in streptozotocin dia-
fatty acid and insulin levels of normal

1292 DIABETES CARE, VOLUME 26, NUMBER 4, APRIL 2003

Yeh and Associates

betic rats treated with an aqueous extract diabetes mellitus. Phytomedicine 3:241– 113. Goyal DK, Tiwari SK: Study on sadaba-
of Ficus carica leaves. Phytother Res 10: 243, 1996 har (Vinca rosea Linn) in madhumeha
82– 86, 1996 101. Ghannam N, Kingston M, Al-Meshaal (diabetes mellitus). J Res Ayurveda Sid-
89. Frati AC, Diaz NX, Altamirano P, Ariza IA, Tariq M, Parman NS, Woodhouse N: dha XX:93–100, 1999
R, Lopez-Ledesma R: The effect of two The antidiabetic activity of aloes: prelim-
sequential doses of Opuntia streptacantha inary clinical and experimental observa- 114. Vray M, Attali JR: Randomized study of
upon glycemia. Arch Invest Med (Mex) tions. Horm Res 24:288 –294, 1986 glibenclamide versus traditional Chi-
22:333–336, 1991 102. Koo ML: Aloe vera: antiulcer and antidi- nese treatment in type 2 diabetic pa-
90. Frati AC, Gordillo BE, Altamirano P, abetic effects. Phytother Res 8:461– 464, tients. Diabetes Metab 21:433– 439,
Ariza CR, Cortes-Franco R, Chavez-Ne- 1994 1995
grete A: Acute hypoglycemic effects of 103. Ajabnoor MA: Effect of aloes on blood
Opuntia streptacantha Lemiare in NIDDM glucose levels in normal and alloxan di- 115. Hale PJ, Horrocks PM, Wright AD,
(Letter). Diabetes Care 13:455– 456, abetic mice. J Ethnopharmacol 28:215– Fitzgerald MG, Nattrass M, Bailey CJ:
1990 220, 1990 Xiaoke tea, a Chinese herbal treatment
91. Frati-Munari AC, Gordillo BE, Alta- 104. Bunyapraphatsara N, Yongchaiyudha S, for diabetes mellitus. Diabet Med 6:675–
mirano P, Ariza CR: Hypoglycemic effect Rungpitarangsi V, Chokechaijaroenporn 676, 1998
of Opuntia streptacantha Lemaire in O: Antidiabetic activity of Aloe vera L
NIDDM. Diabetes Care 11:63– 66, 1988 juice. II Clinical trial in diabetes mellitus 116. Xiong MQ, Liang LW, Lin AZ, Wu QH,
92. Velussi M, Cernigoi AM, De Monte A, patients in combination with gliben- Zhang GL, Miao LP, Zheng JX, Liu JQ:
Dapas F, Caffau C, Zilli M: Long-term clamide. Phytomedicine 3:245–224, 1996 Clinical and experimental study of Per-
(12 months) treatment with an anti-ox- 105. Ni YX, Liu AQ, Gao YF, Wang WH, YG sical Decoction for Purgation with Addi-
idant drug (silymarin) is effective on hy- S, Wang LH, Zhang YH: Therapeutic ef- tion in the treatment of non-insulin
perinsulinemia, exogenous insulin need fect of Berberine on 60 patients with dependent diabetes mellitus. CJIM
and malondialdehyde levels in cirrhotic non-insulin dependent diabetes mellitus 1:96 –99, 1995
diabetic patients. J Hepatol 26:871– 879, and experimental research. CJIM 1:91–
1997 95, 1995 117. Ryan EA, Imes S, Wallace C, Sherman J:
93. Persaud SJ, Al-Majed H, Raman A, Jones 106. Chandola HM, Tripathi SN, Udupa KN: Herbal tea in the treatment of diabetes
PM: Gymnema sylvestre stimulates insu- Hypoglycaemic response of C tamala in mellitus. Clin Invest Med 23:311–317,
lin release in vitro by increasing mem- patients of maturity onset (insulin inde- 2000
brane permeability. J Endocrinol 163: pendent) diabetes. J Res Ayurveda Siddha
207–212, 1999 1:275–290, 1980 118. Tenzin Namdul, Sood A, Ramakrishnan
94. Baskaran K, Ahamath BK, Shanmu- 107. Iyer UM, Mani UV: Studies on the effect L, Pandey RM, Moorthy D: Efficacy of
gasundaram KR, Shanmugasundaram of curry leaves supplementation (Mur- Tibetan medicine as an adjunct in the
ERB: Antidiabetic effect of a leaf extract raya koenigi) on lipid profile, glycated treatment of type 2 diabetes. Diabetes
from Gymnema sylvestre in non-insulin proteins and amino acids in non-insulin Care 24:176 –177, 2001
dependent diabetes mellitus patients. J dependent diabetic patients. Plant Food
Ethnopharmacology 30:295–305, 1990 Hum Nutr 40:275–282, 1990 119. Sircar AR, Ahuja RC, Natu SM, Roy B,
95. Welhinda J, Karunanayake EH, Sheriff 108. Kohli KR, Singh RH: A clinical trial of Sharma HM: Hypoglycemic, hypolipi-
MHR, Jayasinghe KSA: Effect of Mo- jambu (Eugenia jambolana) in non-insu- demic and general beneficial effects of an
mordica charantia on the glucose toler- lin dependant diabetes mellitus. J Res herbal mixture MA-471. Altern Ther Clin
ance in maturity onset diabetes. J Ayurveda Siddha XIV:89, 1993 Pract 3:26 –23, 1996
Ethnopharmacology 17:277–282, 1986 109. Kudolo GB: The effect of a 3-month in-
96. Srivastava Y, Venkatakrishna-Bhatt H, gestion of gingko biloba extract (Egb 120. Chowdhary DP, Dua M, Kishore B,
Verma Y, Venkaiah K: Antidiabetic and 761) on pancreatic beta-cell function in Kishore P: Hypoglycaemic effect of a
adaptogenic properties of Momordica response to glucose loading in individu- coded formulation: Aysuh-82. J Res
charantia extract. Phytother Res 7:285– als with non-insulin-dependent diabetes Ayurveda Siddha XIX:107–111, 1998
289, 1993 mellitus. J Clin Pharmacol 41:600 – 611,
97. Leatherdale BA, Panesar RK, Singh G, 2001 121. Maji D, Singh A: Clinical trial of D-400,
Atkins TW, Bailey CJ, Bignell AH: Im- 110. Moshi MJ, Lutale JKJ, Rimoy GH, Abbas a herbomineral preparation in diabetes
provement in glucose tolerance due to ZG, Josiah RM, Swai ABM: The effect of mellitus. J Diabetic Assoc India 35:1– 4,
Momordica charantia (karela). Br Med J Phyllanthus amarus aqueous extract on 1995
282:1823–1824, 1981 blood glucose in non-insulin dependent
98. Akhtar MS Ahar MA, Yaqub M: Effect of diabetic patients. Phytother Res 15:577– 122. Shankar R, Singhal RK: Clinical assess-
Momordica charantia on blood glucose 580, 2001 ment of the effects of sandana (sandal)
levels of normal and alloxan-diabetic 111. Indian Council for Medical Research: podi-a in the treatment of diabetes mel-
rabbits. Planta Medica 42:205–221, Flexible dose open trial of Vijayasar in litus (neerazhiv). J Res Ayurveda Siddha
1981 cases of newly-diagnosed non-insulin- XV:89: 1994
99. Baldwa VS, Bhandari CM, Pangaria A, dependent diabetes mellitus. Indian
Goyal RK: Clinical trial in patients with J Med Res 108:24 –29, 1998 123. Shankar R, Singhal RK: Clinical studies
diabetes mellitus of an insulin-like com- 112. Iyer UM, Mehta NC, Mani I, Mani U: of the effect of abraga (mica) chendoo-
pound obtained from plant sources. Up- Studies in the effect of dry sundaki (So- ram in the treatment of diabetes mellitus
sala J Med Sci 82:39 – 41, 1977 lanum torvum) powder supplementation (neerazhivu). J Res Ayurveda Siddha XVI:
100. Yongchaiyudha S, Rungpitarangsi V, Bu- on lipid profile, glycated proteins and 108 –111, 1995
nyapraphatsara N, Chokechaijaroen- amino acids in non-insulin dependent
porn O: Antidiabetic activity of Aloe vera diabetic patients. Plant Foods Human 124. Kumar N, Kumar A, Sharma ML: Clini-
L. juice. I. Clinical trial in new cases of Nutr 42:175–118, 1992 cal evaluation of single and herbo-
mineral compound drugs in the man-
agement of madhumeha. J Res Ayurveda
Siddha XX:1–9, 1999

125. Sivaprakasam K, Rao KK, Yasodha R, Ve-
luchamy G: Siddha remedy for diabetes
mellitus. J Res Ayurveda Siddha V:25–32,
1984

126. Lee NA, Reasner CA: Beneficial effect of
chromium supplementation on serum
triglyceride levels in NIDDM. Diabetes
Care 17:1449 –1452, 1994

DIABETES CARE, VOLUME 26, NUMBER 4, APRIL 2003 1293